A Former Employee Reveals Twitter’s Early Efforts To Combat Abuse and Harassment
Share
Roja Bandari, Ph.D., is an Iranian Women’s Rights Advocate and the Co-lead of Women’s March San Francisco, she also worked at Twitter in its early days where she says she worked as new data scientist building a model to detect abuse.
Bandari wrote on Twitter Sunday, “Got a story for you. Years ago when I was a new PhD grad working at Twitter, there was a director of product who had strong ideas about how to solve the abuse and harassment problem on Twitter.”
I worked as a new Data scientist, building a model to detect abuse and then was responsible for A/B testing the solution. HIS product solution, which he was pretty sure was going to change abuser behavior.
Got a story for you. Years ago when I was a new PhD grad working at Twitter, there was a director of product who had strong ideas about how to solve the abuse and harassment problem on Twitter.
— Roja ????️???? (@mathcolorstrees) May 23, 2021
His idea was that if victims did not see the abuse and did not respond to trolls, the trolls would stop the harassment. I did the analysis and no change was detectable in abuser behavior.
I was told this guy had already drafted a press release ready for my numbers to be added, claiming that abuser behavior had changed. My results were not good news. There was real pressure on me to keep looking at different cuts of data. Something not super kosher in stats.
I kept getting the same results. And if anyone asked me, I’d have said his solution wasn’t going to change abuser behavior. Anyone who has ever been harassed online or offline knows that “not engaging” the trolls does not protect you from abuse.
I didn’t have a manager at the time (my new manager was on parental leave). There was no one to push back and be on my side. I remember a Google doc where I kept adding comments saying “no evidence of behavior change” wherever there was such a claim. It killed the press release.
But I felt so powerless, in a team of all men, who had a lot of confidence but very little understanding of the problem and even less empathy for the victims of abuse on Twitter (the Sr eng on the team made some disgusting comments about sexism and racism online).
The team’s director of engineering thought the abuse issue on Twitter was a matter of optics (he said this in a meeting I was in). That it wasn’t that bad and if we quickly shipped a few solutions for the supposed abuse problem the media would move on. I also remember the director of product sitting in a meeting with people from rights groups and folks who were harassed on Twitter and had a very condescending conversation with them. The idea was that he knew what he was doing and these people needed to be handled.
It felt like I was a failure for not getting the “desired” results. What’s worse is that I was part of a wave of layoffs that happened right then, where 30% of employees were let go. I stayed strong but there was a voice inside me that told me I was bad at my job. I’ve been managing teams of engineers and data scientists for a few years now. But the remnants of that first experience have stayed with me. The feeling that I am an outsider. That these arrogant men owned the place even when they did not understand the problem…
Those engineer and product leaders stayed a while and then moved on to the next hot startup and accrued connections and wealth, stabilizing their position in our industry. They didn’t really face any consequences for getting the Twitter abuse problem wrong back then.
I had to get this story off my chest today. But there is a silver lining too.
During that same period I met so many amazing people who worked at Twitter. Coworkers who made me feel connected and whom I enjoyed their company greatly and whom I am friends with to this day
(P.S. I think Twitter has come a long way in terms of cleaning up the timeline from the days of #gamergate. Still would love to see it do better, especially in non-English languages).
